An unsupervised system-call monitoring system using LSTM detects recreated web exploits with AUC up to 0.97 and generally beats PCA and one-class SVM baselines across four applications.
Swaddler: An approach for the anomaly-based detection of state violations in web applications
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
AppMine: Behavioral Analytics for Web Application Vulnerability Detection
An unsupervised system-call monitoring system using LSTM detects recreated web exploits with AUC up to 0.97 and generally beats PCA and one-class SVM baselines across four applications.